From the 7th until the 11th of April, RoMaDS will host Johannes Schmidt-Hieber (University of Twente) with the mini-course

Statistical theory of deep learning

The schedule is as follows: Mon 14:00-16:30, Wed 14:00-16:30, Fri 14:00-16:30.

All lectures will be held in Aula Dal Passo in the Math Department of university of Rome, Tor Vergata.

Here is the program for the three lectures:

Lecture 1. Intro and theory for shallow networks

Perceptron convergence theorem
Universal approximation theorem
Approximation rates for shallow neural networks
Barron spaces

Lecture 2. Theory for deep networks

Advantages of additional hidden layers
Deep ReLU networks
Misclassification error for image deformation models

Lecture 3. Theory of gradient descent in machine learning

Optimization in machine learning
Weight balancing phenomenon
Analysis of dropout
Benign overfitting
Grokking

We encourage in-person partecipation. Should you be unable to come, here is the link to the Teams streaming:

https://teams.microsoft.com/l/meetup-join/19%3arfsL73KX-fw86y1YnXq2nk5VnZFwPU-iIPEmqet8NCg1%40thread.tacv2/1742807097614?context={"Tid"%3a"24c5be2a-d764-40c5-9975-82d08ae47d0e"%2c"Oid"%3a"650fc4a8-4cec-4bd2-87bc-90d134074fe6"}

The seminars are part of the Excellence Project MatMod@TOV.